Flood Forecasting Based on Radar Rainfall Measurements

A hydrometerological study in a river basin was conducted to demonstrate the use of a weather radar for rainfall and flood-flow forecasting. The study region is a Pinioç river subbasin with a drainage area of 2,763 km 2 located in central Greece. Six storm events that created flash floods were recorded every 10–30 min, and were analyzed and processed in terms of ground clutter supression, anomalous propagation, beam refraction, and the losses and merging procedures of radar and rain-gauge rainfall data. A rainfall-runoff model for flood-flow forecasting based on the unit hydrograph theory was then applied twice. First, by using as input the mean areal rain-gauge rainfall derived by applying the Thiessen polygon method and second, by applying the mean areal radar rainfall information over the basin. It was found that the model performs better when it uses processed weather radar data as input. Results of using as input oneand two-hour forecasted rainfall for flood forecasting were encouraging.